Radak Divrei HaYamim Manuscripts

Thirteen known manuscripts have preserved all or part of the Radak’s commentary on Chronicles.  Of these, seven are complete or nearly so, and the rest are fragmentary, containing anywhere from just a few lines to the majority of the work.  The first printed edition, on which all subsequent editions are either directly or indirectly based, serves as an additional text-witness.  Following is a brief description of each text-witness, its relationship to the others, and its role in the edition. The date and provenance of the manuscripts is provided based on analysis conducted by Prof. Malachi Beit-Arié and Dr. Enda Engel. One additional manuscript, Oxford: Bodleian Opp. Add. 125, written in the 1660s in Oriental script, is copied from a printed text and is of no value as a witness.

1. MS Escorial G II 6 (ס) is a complete and legible text, written in semicursive Spanish script from the late 13th to early 14th century.  It is probably the earliest of all witnesses, and is the only one written on parchment.  It serves as the base-text for the edition despite its fair share of errors.  This is because no other witness has preserved a discernibly better text, and this manuscript contains almost all of Radak’s many later additions to the commentary, including a small number—such as a comment on the word ותרשישה at 1 Chr 1:7—that appear in no other witness except one closely related manuscript.  The body of the edition represents the text of MS Escorial (including the rare instances—noted in the critical apparatus—where its text reflects the hand of a later scribe who inserted appropriate corrections), and reproduces its spellings, abbreviations, etc., with the following primary exceptions: (1) where there is overwhelming evidence of error, or (2) where the manuscript is in fact incomplete in its incorporation of later additions.  Such departures from the base-text are marked by asterisks and noted in the apparatus.  Other changes include the spelling of God’s name יי' instead of 'ייי; the presentation of all words representing letters or vowels, such as אלף and פתח, as אל"ף and פת"ח in order to smooth out an inconsistency in the manuscript; and the correction of problematic lemmas (to be discussed below). The sporadic instances of vocalization in the manuscript are not always included in the edition (these do not appear to be original), and other vowels are occasionally added for the purpose of clarification.

2. MS Parma 3569 (א) contains the text of Chronicles with Radak’s commentary in the margins.  The commentary is transcribed in 14th- to early 15th-century semicursive Ashkenazic script, and it contains the entire text until 2 Chr 35:18, which is near the very end.  This manuscript is closest to the base-text, sharing with it the greatest number of errors and containing the same later additions.  However, it suffers from problems of legibility in many instances, including some long stretches, and contains frequent errors resulting from careless transcription.  It appears to preserve no uniquely compelling readings, except in the following rare cases: (1) where the base-text, the only other witness in this branch, developed an error parallel to that of all other witnesses; and (2) where the scribe of MS Parma (or that of a close ancestor)—consciously or otherwise—emended a problematic reading.  For example, at 1 Chr 1:36, most manuscripts, including MS Escorial, contain the problematic reading פסוק, others have פסק (probably the correct reading), while MS Parma contains the plausible reading פסיק, a rare word that could easily have been corrupted to read פסוק in more than one manuscript tradition.  Otherwise, any persuasive reading in this manuscript that differs from that of the base-text will find corroboration in other witnesses.  The edition includes the variants found in this witness only selectively, so as not to swell the apparatus with implausible readings.

3. MS JTS Lutzki 865 (ב), written in 15th-century semicursive Spanish script, attests to a fragment of the commentary, from 1 Chr 1:32 to 1 Chr 26:5.  It is faded, torn in places, and readable only in the original.  This manuscript is also closely related to the base-text, and its limited value is similar to that of MS Parma: in instances where it offers an attractive reading that differs from the base-text, the variant will find corroboration in other witnesses, barring any coincidental parallel development of error, and provided that the attractive reading does not result from a scribal emendation.  For example, the double appearance of the word אשתו in this manuscript at 1 Chr 3:3—found in no other primary witness except as a late insertion in one other manuscript—is in fact a plausible reading, while the omission of one appearance of the word in the other text-witnesses might well have resulted from “haplography” (whereby one of two successive identical words is inadvertantly omitted)—a type of error so common that it could easily have occurred here in multiple manuscript traditions.  Also, the unique, instinctively attractive reading אחרת in this manuscript at 1 Chr 1:32—in place of אחת—probably results from an understandable but apparently misguided scribal emendation.  As with MS Parma, variants of this witness are included selectively, with the omission of its many implausible readings.

4. MS Modena: Archivo di Stato 145 and MS Nonantola: Archivo Comunale 13 contain brief fragments of a common manuscript (ד), the former from 2 Chr 23:5 to 2 Chr 24:7, and the latter from 2 Chr 8:15 to 2 Chr 11:23 and from 2 Chr 34:31 to 2 Chr 36:10.  It is written in 14th-century semicursive Ashkenazic script, and is also closely related to the base-text.  It tends toward shorthand and contains no unique and compelling readings.  Generally, only the variants that it shares with other witnesses are included in the apparatus, for the purpose of showing its relationship to those witnesses.

5. MS Florence: Marucelliana C.CCCLXI (ר) is a complete and legible text written in 14th- to early-15th-century semicursive Spanish script, and is in the family of the base-text. It is the only witness of comparable quality to the base-text, although it appears to be later, and it lacks the small number of Radak’s additions that appear only in the branch of the base-text.  On rare occasions, this manuscript contains suggestive readings that are largely unattested in other witnesses.  For example, at 2 Chr 4:5, it includes—along with just one other erratic witness—the phrase שעמקו חצי רחבו, an important qualification of Radak’s comment that renders it consistent with a parallel comment in his commentary to Kings.  While this particular phrase may be a gloss by someone other than Radak, such examples suggest that there may be instances, however uncommon, where this specific manuscript has preserved valuable variants gleaned from manuscript traditions not otherwise represented in the corpus of witnesses.  The fair quality of the manuscript facilitates the possibilty of comprehensive representation in the apparatus, and this option was chosen to avoid subjectivity in excluding variants.  For this witness, therefore, the absence of a variant in the apparatus implies agreement with the base-text, and any reading may be fully reconstructed based on the information provided.

6. MS JTS Lutzki 784 (ת), written in 14th- to 15th-century semicursive Spanish script, contains brief fragments from 2 Chr 3:15 to 2 Chr 4:5.  The little evidence it presents appears to place it in the family of the base-text.  Its variants are included selectively.

7. MS Paris: National Library Hebrew 198 (פ), written in 15th-century Byzantine script, belongs to a different family of witnesses.  Its most striking feature is that the text in the body of the manuscript preserves the commentary in an earlier form than does any other. Large segments of the commentary that Radak apparently added later—particularly after his work on parallel material in the Former Prophets—appear in this manuscript as marginal and interlinear insertions.  These are also in 15th-century Byzantine script, but in different handwriting.  In the case of the last page, some passages in their later, expanded form appear on top of erased portions of the previous, shorter text.  Because the scribe who incorporated these later additions consulted an independent exemplar, he also uses it to emend the original text of the manuscript.  The source of these additions and emendations appears to have belonged to still another family of manuscripts.

These observations highlight a serious difficulty in placing some of our witnesses in the stemma (that is, the “family tree” of text-witnesses): what if some or all of Radak’s additions were not represented in a certain early manuscript, but were incorporated into its descendants based on an exemplar of different origin—as we can see occurred in the case of MS Paris?  For those descendants, analysis of variants in passages that fall within these later additions might require markedly different stemmatic evaluation.  Furthermore, what if the scribe, in the context of consulting another exemplar for the purpose of including the later material, also emended the text in other respects—again as in the case of MS Paris?  Such corrections might produce an intensely conflated tradition.  Indeed, in the case of MS Escorial and MS Marucelliana, we have already observed the incorporation of a number of later additions, apparently based on independent exemplars.  In the case of that family, however, analysis of shared errors in most of Radak’s additions reveals no unusual distribution, which suggests that the added material was essentially integrated into that tradition early—prior to the emergence of the branches represented by our manuscripts.  For this family, then, conflation is not a serious concern: the various branches reflect one common source, even if that source drew from different traditions for different parts of the work.  Concerning other witnesses, however, we shall not be quite so lucky.

Now despite the division between the Paris manuscript and the base-text family, there is some evidence of an early common ancestor.  For example, at 2 Chr 5:5, the Escorial-Parma branch is missing the same word והארון as is MS Paris.  And although this glaring omission does not appear in other manuscripts in the two families (probably due to scribal correction), it is difficult to ascribe to coincidence the appearance of this error in any two witnesses.  Also, at 1 Chr 1:36, instead of what should read הנראה, the base-text family has the untenable בנראה and MS Paris has what appears to be the same reading, or perhaps the similar כנראה which reads smoothly but is inconsistent with Radak’s usage.  Finally, at 1 Chr 1:38, what should clearly read כי appears in the base-text family as the untenable וכי, which seems to be related to the equally problematic reading וכן in MS Paris. This probable common ancestry bears important implications. In cases of conflict between manuscripts, we would generally conclude that a reading that appears in more that one family is correct, unless there is good reason to suspect that the same error could have developed coincidentally in two independent traditions.  If there is common ancestry between two families, however, we cannot favor a reading common to both on these grounds alone.  Thus, a reading need not be considered to be decisively accurate just because it is common to MS Paris and the base-text family, as they appear to share a such a common ancestor.

Finally, on the matter of representation of variants of MS Paris in the apparatus, we again confront the problem of an overwhelming number of errors.  This is compounded by the invasion of the manuscript by the later scribe, whose emendations make the original text occasionally indecipherable.  There is the additional matter of representation of the later insertions.  For a complete witness as important as MS Paris, complete representation would have been preferable.  The problem with excluding variants becomes particularly acute when one considers the matter of common ancestry, which makes it more difficult to favor one reading over another.  While it nevertheless proved unwise incorporate all the variants of MS Paris, considerable caution has been exercised in excluding its problematic readings, in an effort to represent as faithfully as possible this early version of the commentary. Incorporation of the insertions of the later scribe follow the same policy.

8. The first printed edition (editio princeps) of the commentary (נ) appears in the 1548 Venice Mikraot Gedolot.  All later editions follow the basic text of the editio princeps, and it appears that variations between editions generally result from conjectural emendation.  Such attemped corrections are most ubiquitous in the Lublin Mikraot Gedolot, which promises to clean up errors, as well as in recently published editions.  The relatively late publication of all of these, as well as the quality of the corrections in light of manuscript evidence, strongly suggest that the emendations are entirely speculative.

The editio princeps is a close descendant of MS Paris (with the insertions), as a number of its errors are indisputably attrributable to misreadings of that manuscript.  Almost all of MS Paris’s errors reappear in the editio princeps, while the latter adds a number of its own.  While there are occasional readings in the editio princeps that seem to improve upon MS Paris, it is not clear whether this results from conjecture or from sporadic consultation with another manuscript.  Variants of the editio princeps are incorporated systematically.

9. MS Munich Hebrew 363 (מ), written in semicursive Spanish script from the late 13th to early 14th century and mostly legible, preserves an early text of the commentary, similar to the body of MS Paris.  It does, however, contain a few of Radak’s later additions.  Many minor errors in this manuscript—omitted words, more often than not—have been corrected by a later scribe.  Beyond their common preservation of an earlier version of the commentary, MS Munich and MS Paris appear to be stemmatically related, based on a handful of shared errors that are hard to ascribe to coincidence, such as the problematic omission of the lemma at 2 Chr 20:9.

Notably, there also appear to be instances of shared error between MS Munich and the base-text family.  For example, they both contain the problematic omission of הם בני הגר at I 5:10, despite its appearance in MS Paris.  Furthermore, in the cases where we noted shared errors in MS Paris and the base-text family, MS Munich contains similar errors, with the exception of the omission of והארון at 2 Chr 5:5, where MS Munich—like some representatives of the base-text family itself—might have undergone later correction.  As with MS Paris, then, it appears that there is not a perfectly sharp dichotomy between MS Munich and the manuscripts closely related to the base-text.  Also, the incorporation of some later additions suggests that the tradition represented by MS Munich might have experienced contamination at some point.  As noted, the less we can isolate fully independent manuscript traditions, the harder it becomes to outnumber and exclude variants.  To avoid subjectivity in evaluating potentially meaningful variants, the Munich manuscipt is represented in the apparatus fully, despite its large number of errors.

10. MS Vatican Hebrew 89 (ו) is written in 14th-century semicursive Spanish script, and is mostly legible.  It appears, based on a few decisive common errors (and on the cumulative evidence of some less decisive ones), to share an ancestor with the base-text family.  Consider, for example, the problematic לגלות instead of ולגלות at 1 Chr 4:27-33, the apparently indefensible omission of של עשו at 1 Chr 4:42, and the omission of הם בני הגר at 1 Chr 5:10.  On the other hand, MS Vatican also manifests a close relationship to the insertions in MS Paris, and the two include a number of later additions not found in any other witness (with the exception of the editio princeps, which is a descendant of MS Paris).  However, since most of the evidence for a relationship between MS Vatican and the Paris insertions falls within the later additions, which comprise the vast majority of these insertions, we might argue that the connection is in fact limited to the later material.  It is tempting, in fact, to view the majority of MS Vatican as a relative of the base-text, and to place it in the family of the Paris insertions for its representation of the later additions.

However, even within passages that are probably not later additions to the commentary—for they appear in the body of MS Paris—MS Vatican shares certain errors with the later scribe’s emendations of MS Paris.  For example, at 2 Chr 9:12, there is only a small possibility of coincidence for the appearance of the problematic ונתן—in place of נתן—in precisely these witnesses (and their close relatives).  This suggests that the relationship between MS Vatican and the Paris insertions extends even to passages that derive from the commentary in its earlier form.  Also, in what is a later addition, at 1 Chr 24:3, MS Vatican shares the problematic שנים—instead of שוים—with the base-text family, while the Paris insertion has the correct reading.  While this last anomaly might be a coincidence, there still emerges considerable difficultly in the attempt to draw a neat stemmatic distinction within MS Vatican between its representation of the earlier and later material, and we must seriously consider the possibility, even the probability, of conflation.

For all its similarities to other manuscripts, MS Vatican nonetheless retains many unique readings.  For the most part, however, they consist of careless errors or reflect significant liberties taken by the scribe, including truncation, shorthand, word transpositions, attempted corrections, emendation of elliptical style, and replacement of Radak’s terminology with more pedestrian synonyms (“trivialization”), such as כאן for הֵנָּה.  On some occasions, there are genuinely plausible unique variants, such as בסבתא וסבתה at 1 Chr 1:7 and the addition of כלו at 1 Chr 3:1.  As with MS Paris, full representation of this important manuscript would have been preferable.  However, even more than in the case of MS Paris, this would have swelled the apparatus with far too many “nonsense variants.”  Therefore, cautiously but regularly, implausible variants in MS Vatican have been omitted.

11. MS Strasbourg 4028 (ג), written in 14th-century semicursive Spanish script, contains a few lines, from 1 Chr 4:23 to 1 Chr 4:33.  It seems to be rather close to MS Vatican, as evidenced by the selected variants included in the apparatus.

12. MS St. Petersburg: Russian National Library II A 6 (ט), written in 16th-century semicursive Byzantine script, contains two brief sections of the commentary, from the beginning until 1 Chr 4:22 and from 2 Chr 24:11 until the end.  For the later additions, this manuscript is related to MS Vatican and the Paris insertions.  However, in one crucial instance, at 1 Chr 3:8, where these three are the only witnesses (other than the editio princeps) to a certain addition, MS St. Petersburg is the only manuscript to insert it in the proper place.  This, in addition to some other errors shared by the other two that do not appear in MS St. Petersburg, places this manuscript in its own branch for the later additions.

For the rest of the commentary, in instances where we have previously seen evidence of an overarching relationship between the base-text family and the Paris-Munich family, at 1 Chr 1:36 and 1 Chr 1:38, MS St. Petersburg shares the same errors, in the latter case against the reading of a Paris insertion. (The evidence of MS Vatican is unclear in these cases.)  On the whole, the precise place of this manuscript in the stemma is not entirely clear, and it is, once again, difficult to reject its unique readings solely on the basis of stemmatic analysis (except in the context of the later additions, where its place is clear).  Due to its many errors, however, its apparently implausible variants are omitted.

13. MS Oxford: Bodleian Opp. Add. folio 24 (ק) contains the commentary of Rabbi Benjamin ben Judah of Rome to Chronicles, with a mostly legible text of Radak on the margins in 15th-century Italian script.  For Radak, the manuscript contains from 1 Chr 1:1 to 1 Chr 3:10 and from 2 Chr 23:4 until the end.  In instances where R. Benjamin addresses a problem in a manner similar to Radak, however concisely, Radak’s parallel comment, however elaborate and multifaceted, is generally omitted.  A later scribe filled in much of the gap in the form of marginal and interlinear insertions.

For later additions, this text is related to MS Vatican, as evidenced, for example, by their shared error לא הועיל at 2 Chr 28:20.  (Notably, in MS Oxford, there is an overline here, which apparently serves to indicate that this might be an error.)  In other cases, particularly in the context of the additions, MS Oxford contains some striking expansions; for example, at 2 Chr 24:27, where other witnesses contain a mere allusion to a prior comment, MS Oxford contains a lengthy explication.  It is, however, unclear if such expansions are from the pen of Radak.  For the commentary as a whole, this witness contains some unique and plausible readings.  Its place in the stemma is hard to trace based on the available evidence, although it might bear a resemblance to MS Vatican and/or MS St. Petersburg.  The matter of legibility complicates the representation of this witness in many instances, but its meaningful variants are included as fully as possible.

In summary, MS Escorial serves as the base-text for the edition.  The critical apparatus fully represents MS Marucelliana, MS Munich, and the editio princeps, and presents selective variants of the others.  The stemmatic relationships may be presented as follows:

Radak DHY MSS Diagram

In this stemma, (A) represents the Urtext; (B) the likely archetype responsible for errors represented in all families, such as the omission of הם בני הגר at 1 Chr 5:10; and (C), (D) and (E) the common originals of the three families.  (H) represents the branch of the base-text (ס).  The placement of MS Oxford (ק), MS St. Petersburg (ט), the Paris insertions (פ1) and MS Vatican (ו) in the same family is speculative for the core of the commentary, but conclusive for the later additions.  The arrow shows the insertion of the later additions into MS Paris (פ).  We must also bear in mind a relationship between MS Vatican and the base-text family that results from either conflation or a closer common ancestor than is represented here.  Other instances of conflation, such as the inclusion of some extra additions into the base-text branch based on an independent tradition, are also not represented.  Finally, the layout should not leave the impression that the base-text is particularly far-removed from the Urtext.  Rather, there are simply a large number of  extant relatives of the base-text, and their representation in the stemma nudges this relatively early, full, and accurate text to the bottom of the picture.

This concludes our discussion of primary witnesses.  There are also two notable secondary witnesses: the 16th-century commentary of R. Solomon ibn Melekh, Mikhlal Yofi (ל), most of which is verbatim quotation of Radak, and the 13th-century commentary Midrash Divrei Ha-Yamim le-Rabbi Shemu’el Masnut (ש), available in full only in MS Vatican Hebrew 97, which has as its basis regular verbatim quotations of Radak.  (Editions that reproduce the layout of the Lublin Mikraot Gedolot and contain Malbim rather than ivre teitch at the bottom contain the first few chapters of Masnut, also at the bottom, beginning on the page where 2 Chr 9 begins.) Both of these commentators appear to have utilized texts related to MS Paris, and R. Samuel Masnut clearly lacked the later additions.  Variants from these witnesses appear in the apparatus in parentheses where they might present useful evidence, particularly when they support an emendation made to the base-text.  Of particular interest is the variant חכמים at 1 Chr 28:21 in R. Samuel Masnut, probably the correct reading, in the midst of a line that is omitted in multiple witnesses due to “homoioteleuton” (whereby the scribe’s eye skips from one occurrence of a word to a later one).

A different policy obtains for representation of variant readings of lemmas.  Erratic transcription of lemmas by scribes, inconsistencies in marking them as lemmas, the imposition of plene spellings, and attempts—particularly in the editio princeps—to clean things up and present lemmas accurately have all contributed to a special problem for their representation both in the body of the edition and in the apparatus.  The following approach has been taken.  For plene and defective spellings, lemmas follow the readings in M. Breuer’s edition of the Bible, because the text-witnesses are so unreliable in this respect, generally reflecting no more than scribal idiosyncrasies.  Otherwise, where the lemma in the base-text is entirely defensible, it is reproduced as is, and no variants are provided in the apparatus unless they are similarly defensible and important for the sense of Radak’s comment.  Where a reading in another witness is actually more persuasive than that of the base-text (such as the occasionally important addition of וכו'), it is incorporated into the body of the edition.  In such cases, the apparatus cites at least one manuscript that serves as a basis for the emendation, followed by the base-text’s reading.  Where the lemma in the base-text deviates from all variants of the biblical text listed in the collections of J.B. de-Rossi and C.D. Ginsburg, the text found in Breuer’s edition is followed, and the apparatus provides a supporting variant from at least one manuscript of Radak where available, as well as the base-text reading.  The “one manuscript” chosen in these cases is generally MS Marucelliana, because its departure from the base-text in such instances shows how the better reading is preserved in the latter’s very own family.  Where MS Marucelliana supports the problematic base-text reading, the reading MS Paris is generally chosen, in order to demonstrate how evidence from another family conflicts with these decisive representatives of the base-text family.  Where a variant of the biblical text attested in the above collections supports the base-text reading, against the reading in Breuer, the base-text reading is retained in the body of the edition, and Breuer’s reading appears in the apparatus.  The reader might wish to note several instances where the base-text, usually along with other witnesses, suggests that Radak had before him a reading of the biblical text that is attested in only a small minority of witnesses cited by de-Rossi and/or Ginsburg.  These include: ולא היו at 1 Chr 2:34, צדקיהו at 1 Chr 3:16, ומלכי רם at 1 Chr 3:18, אבי at 1 Chr 4:11, משפחותם at 1 Chr 4:27, אביאסף at 1 Chr 6:7, ובהשחית at 1 Chr 21:15, ומעשֶה at 1 Chr 23:27, and אביחיל at 2 Chr 18:11.

On the other hand, for biblical citations in the text of Radak’s comments, even unsupported deviations from the Masoretic text have generally been left intact, because Radak evidently often cited from memory and presented texts imprecisely (see especially at 1 Chr 18:4).  Where other witnesses conflict with the base-text in this regard, standard policies of representation apply.

Certain inconsequential variants have been systematically omitted from the apparatus. These include plene and defective spellings; other differences in spelling, such as יחש/יחס; variants resulting from abbreviation; common and irrelevant variations such as אחר/אחרי‎, ‏וגו'/וכו'‏‎, ‏אלו/אלה‎, ‏כן/כך‎ and אינה/איננה; and differences in verb forms such as אומר/אמר—generally where their subject is the biblical author—in cases where the distinction is of no consequence.

As noted, the branches of the base-text family (‏ס‎, א‎, ב‎, ד‎, ר‎, and ת) seem to have remained largely uncontaminated after their emergence.  As a result, in cases where branches other than that of the base-text agree with another family, the base-text reading can generally be considered outnumbered by representatives of distinct traditions.  Nevertheless, where there are no other grounds to assume that the base-text is in error, considerable caution has been exercised before emending its text in the body of the edition.  This is because of the small amount of contamination that apparently did take place, the possibility of coincidental parallel development of error, and the systemic value of a conservative approach to the emendation of a base-text.

Finally, the reader who wishes to identify quickly examples of later additions not found in the body of MS Paris or in MS Munich is advised to locate the many instances of “מפ ח' (פ1 השלים)” in the apparatus.  These will only occasionally reflect errors of omission, and usually indicate material that did not appear in earlier versions of the commentary.  For additions that MS Munich has incorporated, the apparatus will generally read “פ ח' (פ1 השלים).”